In this study, the authors test two types of practice combined with two types of patterns. The patterns are the expanding and uniform space patterns mentioned before. And the types of practice are: **Repetition practice:** The subjects are presented with the same information at different practice intervals. **Test-type practice:** The subjects are presented with information once and then tested on that information at different practice intervals. If you're interested in this topic, I highly recommend Gwern's review: https://www.gwern.net/Spaced%20repetition This plot shows the forgetting curves for the test-type practice of Experiment 2. We can clearly see that the expanding space pattern gives better results than the uniformly spaced pattern (with the same average spacing). The final test shows a **substantial improvement in retention relative to the test group of 2.5 times better retention for the uniform test pattern and 3 times for the expanding test pattern.** This plot shows the forgetting curves for the different types of rehearsal patterns during the study phase. We can see that **the expanded space pattern ((0,3,10) and (1,4,10)) gives better retention results than the uniform and the contracting space patterns.** It is interesting to notice the difference in short term forgetting-difference between the expanded patterns (0,3,10) and (1,4,10). **The shorter the initial interval the better for longer term retention.** A final retention test was taken 30 min after the study phase. This plot shows the retention results and relative improvement for the different learning patterns. **The expanding space pattern is substantially better than the uniform and contracting space pattern**. The expanding space pattern **yields almost 2 times better retention when compared to the presentation alone, where the subjects can their own rehearsal strategies.** It is also interesting to notice that the **short term uniform pattern** (0,0,0) although good for short term retention **is the worst pattern for longer term retention.** The results for the final test show that the **expanding-test pattern** is the best technique for longer term retention. The results for the **final retention - proportion recalled - the four technique-pattern combinations are presented below.**  **Robert A. Bjork** holds a PhD in Psychology from Stanford. He is Distinguished Research Professor in the Department of Psychology at the University of California, Los Angeles and a Fellow of the American Academy of Arts and Sciences. His research focuses on human learning and memory and on the implications of the science of learning for instruction and training.  **Test-type learning** is a type of learning where information is presented just once and then is subsequently tested. A good example of this type of learning is name-learning. You are presented with the name of a person once and then do self-administered tests until you learn to associate the name with the person. The authors suggest that when using test-type practice it is best to **use expanding time intervals between tests** instead of uniform time intervals in order to optimize longer term retention. This type of learning technique is called [Spaced Repition](https://en.wikipedia.org/wiki/Spaced_repetition). Watch the video below to learn more about spaced repetition in learning theory. [](https://www.youtube.com/watch?v=cVf38y07cfk) **Thomas K. Landauer is one of the pioneers of Latent Semantic Analysis**, a technique of Natural Language Processing which mathematically models people's knowledge of words and their semantic interrelationships. He completed his PhD at Harvard University in 1960 and became a psychology professor and taught at Dartmouth, Stanford, Princeton, and the University of Colorado (where he was an emeritus professor). In 1969 he took a job at Bell Laboratories in the basic research area where **he began one of the first research groups in human-computer interaction.** He made major contributions to experimental psychology, anthropology, and cognitive science.  The subjects of this study were presented with cards containing **first and last names of fictitious characters**. In the initial presentation they were given both names and in the next three presentations they were given the first name and tested on the last name. The timed intervals differed from: - **Uniform:** containing a fixed number of events between the successive presentations - **Expanding:** containing an increasing number of events between the successive presentations, i.e. 0 events, 3 events, 10 events. - **Contracting:** containing an increasing number of events between the successive presentations, i.e. 10 events, 3 events, 0 events. The **Forgetting Curve** is the result of a simple experiment that was the first to demonstrate that there is an exponential loss of memory unless information is rehearsed over time. The forgetting curve can be roughly described by: $$ R=e^{-{\frac{t}{S}}}$$ $R$ is memory retention, $S$ is the relative strength of memory, and $t$ is time. It was first introduced by [Hermann Ebbinghaus](https://en.wikipedia.org/wiki/Hermann_Ebbinghaus) in 1885. Ebbinghaus memorized a series of nonsense syllables and then tested his memory of them at various periods ranging from 20 minutes to 31 days.  This paper shows how spaced learning can influence the Forgetting Curve and increase longer term retention.